Red Hat Needs to Lighten Up

When I learned that Red Hat Enterprise Linux 5, which is a big release for Red Hat to which I’ve been looking forward for some time, was coming out on March 14, one of the first thoughts that crossed my mind was, “Great, when’s CentOS 5 coming out?” Even though Red Hat has always been very nice about providing us with entitlements to test their products, entitlements are a major pain to mess with. Sometimes our entitlements expire, and I have to head over to the Red Hat Network to unentitle some machines in order to entitle others– frankly, from our perspective as a testing organization, and as a group of people who often build up and tear down systems in different combinations, RHEL is actually more of a pain to work with than is Windows, for which there were (until Vista, at least) volume license copies that we could use flexibly and without expiration. Even better are free Linux distributions, which you can get in all sorts of forms and from all sorts of locations. Debian is my favorite example of deployment flexibility: I download a small netinstall image, from which I boot a virtual or physical system, choose a network mirror that’s close to me and pull down just the packages I need, in their up-to-date form. Just because Debian is much more pleasant to deploy, however, doesn’t mean that I get to ignore RHEL, which is probably the most important Linux distribution around, in terms of hardware and software certifications and in terms of its prominence among the enterprise infrastructures that our readers are running. Fortunately, there’s CentOS, an open source project that takes the source RPMs that Red Hat diligently offers up for public download, strips out Red Hat’s trademark-encumbered artwork, and improves–significantly–on RHEL by returning to it the flexibility that free distributions like Debian enjoy. The big drawback to CentOS, however, is that while CentOS, practically speaking, really is RHEL, CentOS isn’t RHEL enough for Red Hat to support or furnish services for it. What’s more, running RHEL by some other name puts you in an unclear support situation with ISVs who’ve certified their products for RHEL–just ask Oracle, which in recent weeks has been expressing consternation over certification and its own RHEL rebrand. Back when Red Hat first divided its free, support-optional Red Hat Linux product into the free, bleeding edge and community-supported Fedora and the metered, stable, and Red Hat backed RHEL, it probably made good business sense to bid adieu to any customers unwilling to pay per system. If you wanted a Red Hat distribution with a long support term and a stable development arc, you had no other choice–whether or not you planned on consuming the support for which you’re paying when you buy Red Hat’s free software. However, now that RHEL may be had for free in the form of CentOS, does it really make sense for Red Hat to maintain its self-imposed separation from customers who want a support-optional way to run RHEL? While it’d certainly make life easier for me and for others with needs similar to mine if Red Hat let itself loosen up again, I contend that it’d be Red Hat itself that’d stand to benefit best from the move. For one thing, by allowing CentOS to stand between itself and a growing segment of its user community, Red Hat is missing out on important feedback, bug reporting, and mindshare, which may sound fluffy, but it’s the stuff of which Red Hat’s dominant status in the Linux world was built. Also, Red Hat is allowing its brand to become watered down–as I mentioned above, there’s currently uncertainty regarding the support status of rebranded versions of RHEL. However, as time goes on, and rebrands like CentOS and Oracle’s Unbreakable Linux prove themselves to be truly compatible with RHEL, it’s hard to imagine ISVs turning down the dollars of companies running these clones. Speaking of turning down dollars, Red Hat’s decision to cede a growing portion of its market to CentOS means closing doors to services money from customers who want to run RHEL, but do so with more of the flexibility to which free software is heir. To those who counter that Red Hat can’t stay afloat without requiring all RHEL users to pay for support contracts, whether they want them or not, I say that if Red Hat support delivers real value, then Red Hat has nothing to worry about. If, however, Red Hat’s health truly relies on leveraging customers to buy something they don’t actually need, then the Linux giant is destined for a fall.